Re: Anyone interested a addon plugin that enable you to use 7000 plugins on modern digital scope
Here are some update of what has been happening on this project.
New photos has been uploaded to https://groups.io/g/TekScopes/album?id=246549
Since the initial post, I have made two additional major revisions to the design and board routing. The adapter is in a functional state and also had some extra functionality.
1. Previously the adapter has only one signal output, it has two now. The primary target use of two signal output is to use 7A26's two inputs at the same times. On the original 7000 scope, the second signal output is routed to trigger. On this adapter, you can take the trigger output and connect to another input of the digital scope, this way you will only need two 7A26 to track four input signals. The reason to bring 7A26 into the scheme is that 7A26 has somewhere about 150uV noise floor, which was not easily accessible on CRT display as the maximum sensitivity is limited to 5mV/div.
2. The switching regulator is change to use discrete transistor to overcome overheat issue on the original LT chip.
3. Majority of the parts are move to the bottom side of the board, for mechanical fit reasons.
4. A protective cover PCB is part of the assembly not, so that the board are exposed less to human fingers. +/-50V by itself is boardline safe voltage, but the differential is 100 V, and this is getting out of my comfort zone. Still, the linear regulator transistor's tab has +/-60V on them, and you can still touch them with your finger.
1. The voltage regulation loop is not populated, so the switching regulator runs at constant 50% duty cycle. Thus, the output voltage is more than 60V, and linear regulator has to dump a lot of heat on the heatsink.
2. The linear regulators will overheat if run too long. This can be resolved easily when closed loop voltage control done in #1, the heat dissipation will drop significantly.
3. The snubber circuit is missing on the switching regulator portion, thus when switch turn off, there is a 20 MHz overshot on the switching waveform. This overshot signal somehow finds it's way to the plugins and introduced additional noise. The 7A26's 150uV noise floor is now ruined. The best observed noise floor of the whole signal chain is about 500uV at full bandwidth (3x worse). I am 90% certain the noise comes from the plugins, not from the differential to single-ended conversion amplifier on my adapter, as I can get very good noise floor without insert any plugins (basically measuring the two 50 Ohms termination resistors). I have tried a few combination of snubber circuit, none of them worked.
4. For 7A13, enabling 5 MHz bandwidth filter actually make noise worse, rather than better (self resonant of the internal filters?) .
Is anyone interested in the current state of the adapter ?
At this point, I will need some help from professional power electronics engineers to fix the noise issue. Is any one willing to help? I can provide the schematic privately.
Some basic background information on the DC DC conversion
I used LM25037 to drive a hex-path transformer Coilcraft HPH5-0155. The primary takes two wingdings. The other four are arranged together to provide +/-20 and +/-60V? For the +/-20 output, it is typical push-pull configuration. For the +/-60 it is similar to forward topology. A schematic can be found here https://groups.io/g/TekScopes/album?id=246549
I am willing to change the architecture is some good advice can be given.